In-Situ Extraction of Space Morphology of Electrical Tree Channel in Silicone Gel

This work uses the laser scanning confocal microscope (LSCM) to propose a method for the 3-D fluorescence imaging of electrical trees inside silicone gel, which is used as an insulating material for packaging in power electronic devices. It induces auto-fluorescence in the electrical trees in the silicone gel through laser light to reveal their micro-morphology. The laser light source focuses on the damaged area, and points on different focal planes are successively scanned by using the confocal microscope for the 3-D in situ fluorescence imaging of the electrical trees. Automatic path regression is used to reconstruct the area of the channel of microscopic damage inside the material in situ. The global and local parameters of the electrical trees are statistically calculated from the obtained images to quantitatively analyze their morphology. The calculation found that the diameter of electrical trees was mainly distributed between 4.47–
$32.27~\mu \text{m}$
, and the average diameter of electric branch channels decreased with the increase of voltage frequency. The proposed method provides a novel and feasible technical means for the non-destructive detection of electrical trees inside materials.